OGC P39 The accuracy of Hospital Acquired Pneumonia (HAP) Diagnosis after Oesophagectomy

Abstract

Abstract Background Oesophagectomy is a high-risk intervention associated with significant rates of morbidity. Pneumonia is a complex spectrum of infective and non-infective aetiology, varying in severity and is common after oesophagectomy. Inconsistent definitions of pulmonary complications pose significant challenges in both the clinical and research setting. The internal Esophageal Consensus Complications Group (ECCG) have adopted the Centre for Disease Control (CDC) criteria for HAP diagnosis to address this. However, other validated diagnostic criteria are also available. This study aimed to better understand the incidence of HAP after oesophagectomy, and the accuracy of four validated diagnostic tools. Methods This was a retrospective review of all consecutive oesophagectomies undertaken at a regional OG centre serving a population of 3.5 million patients between 2014-2022. Datapoints included patient demographics; comorbidities; post-operative course and re-admissions. The timing of pneumonia diagnosis was determined in the post-operative period, alongside observations and Computed Tomography (CT) and Chest X-Ray (CXR) reports. Four diagnostic criteria (Centres for Disease control- CDC, American Thoracic Society - ATS, Utrecht and Clinical Pulmonary Infection Score - CPIS) were applied to determine the incidence of HAP in comparison to patients treated for HAP. Statistical tests were performed in STATA. Results 460 patients were included, 356 (77.40%) were male. 223 (48.5%) were treated for HAP between 2014-2022. 56 (12.2%) met the CDC criteria. Mean days to diagnosis was 5 days (SD +/- 3.72). Diagnostic sensitivity and specificity for HAP was 88.89%, and 50.15% respectively. The ATS criteria were met by 79 patients (17.2%), showing a similar sensitivity (97.53%) and specificity (92.02%) to the CDC criteria. The Utrecht criteria had the lowest specificity (75.20%) for HAP diagnosis. 192 patients had sputum samples; 158 (82.29%) were culture positive. 52 patients were re-intubated, of which 29 (55.8%) were reintubated due to respiratory failure. Conclusions HAP is a complex and nuanced diagnosis with high inter-observer variation. Currently available and validated diagnostic tools result in different rates of HAP diagnosis. These findings highlight the importance of adopting robust approaches to diagnosing HAP to avoid unnecessary treatments and to more accurately determine the effectiveness of protocolised management (such as ERAS) that target HAP. The findings also highlight important implications on outcome reporting in research, particularly with interventions (e.g. minimally invasive and robotic surgery) which aim to reduce the burden of post-operative morbidity. Further widescale studies are required to validate these findings.